Skip to content

TDT4173 - Modern Machine Learning in Practice. Predict future vessel positions up to 5 days ahead.

Notifications You must be signed in to change notification settings

andrejvi/AIS-Vessel-Position-Prediction-ML

Repository files navigation

AIS Vessel Position Prediction

TDT4173 - Moderne maskinlæring i praksis

Markus Sandnes
André Joseph Virani Peder Aas Vårheim

Overview

This project is part of the TDT4173 - Modern Machine Learning in Practice course. The main objective is to predict the future positions of vessels based on Automatic Identification System (AIS) data from January 1st, 2024 to May 7th, 2024. The task involves building a machine learning model capable of forecasting vessel positions for five days into the future based on a range of factors such as congestion, port calls, and other relevant events affecting the vessels’ journey.

File overview:

Project description - a brief introduction to the task
Dataset definitions and explanation - a documents that gives more details about the dataset and column names
Training data - training data
Test input data - test input data
Optional data - optional data that contains schedules for some vessels
Optional data - optional data related to ports
Optional data - optional data related to vessels
Example submission - a demo submission to Kaggle what predicts all zeroes
Jupiter Notebook demo - a demo utility function for visualizing the trajectory of a vessel
Score function - the score function we use in the Kaggle competition

About

TDT4173 - Modern Machine Learning in Practice. Predict future vessel positions up to 5 days ahead.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published